accepted manuscript structuring the context for construction … · 2020. 4. 2. · 91 92 2.1 key...
TRANSCRIPT
-
Accepted manuscript
Structuring the context for construction robot development
through integrated scenario approach
Mi Pan, Thomas Linner, Wei Pan, Hui-min Cheng, Thomas
Bock
DOI: https://doi.org/10.1016/j.autcon.2020.103174
To appear in: Automation in Construction
Received 9 October 2019, Revised 3 March 2020, Accepted 7 March 2020, Available
online 31 March 2020.
Please cite this article as: Pan, M., Linner, T., Pan, W., Cheng, H.M. and Bock, T.
(2020). Structuring the context for construction robot development through integrated
scenario approach. Automation in Construction, 114, 103174.
https://doi.org/10.1016/j.autcon.2020.103174.
This is a PDF file of an article of the accepted version. This version will undergo
additional copyediting, typesetting and review before it is published in its final form.
https://doi.org/10.1016/j.autcon.2020.103174https://doi.org/10.1016/j.autcon.2020.103174
-
Accepted manuscript
1
Structuring the Context for Construction Robot Development 1
through Integrated Scenario Approach 2
Mi Pan a, *, Thomas Linner b, Wei Pan a, Huimin Cheng b, Thomas Bock b. 3
a Department of Civil Engineering, The University of Hong Kong, Pokfulam, Hong Kong 4
b Chair of Building Realization and Robotics, Technical University of Munich, 80333 5
Munich, Germany 6
7
Abstract: The technological development of construction robots is underway globally. 8
However, current development activities face significant uncertainties, particularly in 9
terms of the definition and management of system requirements, which are primarily 10
based on vague assumptions about the future. Thus, a new tool is required to grasp how 11
construction robots—and their surrounding ecosystems—will be used. This research 12
adopts an unprecedented scenario-based approach to develop and analyze future 13
alternatives for construction robotics in a systematic manner. Hong Kong “toward 2035” 14
is used as an initial test case, and four scenarios of the robot ecosystem, i.e., “Bottleneck,” 15
“Age of Iron Worker,” “Dynamic Co-evolution of Robotization and Modularization,” and 16
“Rise of the Robots,” are developed from evidence-based analysis. Scenarios highlight 17
the crucial role of workers for construction robot utilization. Driving forces, opportunities, 18
and challenges are identified for elaborating strategies under each scenario. The 19
integrated scenario approach and findings lay an important foundation for systems 20
engineering processes in construction robotics to develop a new tool for structuring 21
system context and specifying system requirements. 22
23
Keywords: construction robotics; construction technology; technology forecasting; 24
scenario techniques; systems engineering. 25
26
* Corresponding author.
Email Address: [email protected] (Mi Pan).
mailto:[email protected]
-
Accepted manuscript
2
1 Introduction 27
The building and construction industry is facing increasingly grave challenges such 28
as cost escalation, skill mismatch, and an aging workforce. This trend is particularly 29
exacerbated in high-tech-focused, high-density, and high-wage metropolitan areas, such 30
as Hong Kong, where the industry is struggling to satisfy the ever-increasing demand for 31
construction [1]. In such contexts, the ability of conventional construction methods to 32
cope with the growing complexity of construction and meet the associated demands for 33
productivity, quality, safety, and sustainability have reached their limits [2]. Construction 34
robotics is considered a promising innovation to address this challenge and reform the 35
industry, which has triggered a plethora of global research and development (R&D) 36
efforts for decades [3]. Many prototypes have been developed since the 1980s, providing 37
evidence on the ability of construction robots to assist in conducting dangerous, 38
monotonous, or tedious construction jobs more efficiently and accurately [4]. 39
Although construction robots promise many benefits, multiple barriers continue to 40
hinder real-world applications. Examples include professional skill shortages; risk-41
averse mentality coupled with industrial potential of complex technology awareness [5]; 42
cultural resistance against the adoption of innovative technologies [6]; and high capital 43
costs and additional setup time of robotic equipment [4]. These barriers lead to 44
uncertainties and eventually allow many possible development lines for next-generation 45
robotic construction. Previous researchers have attempted to predict the future use and 46
utilization of construction robots [4, 7]. However, there is still a lack of systematic 47
exploration of possible future scenarios with multiple alternatives from a systems 48
perspective. In addition, none of these studies have holistically considered the 49
sociotechnical context of construction robots that would surround and determine the 50
-
Accepted manuscript
3
application of such systems. The lack of such studies poses a major challenge for both the 51
early stage conception of construction robots and the associated systems engineering 52
processes [8]. This is because key development aspects (e.g., requirements engineering, 53
business strategies, and degree of automation) that set the stage for the development of 54
construction robotic systems are based on vague assumptions about how the future 55
industry will be and what kind of construction robots will be developed. This situation is 56
aggravated by the fact that construction robotics is not yet an established field, lacking 57
implicit design experiences or robust data from previous developments and applications. 58
Therefore, it is essential to develop new tools to project the future solution space (the set 59
of all possible solutions) of construction robots over a given period. 60
This paper aims to develop a new tool for the systematic development, exploration, 61
and analysis of the future utilization of construction robots in a given industrial 62
environment (ecosystem) based on the application of a scenario approach. Scenario 63
approaches allow users to make assumptions about the mid- to long-term future [9, 10], 64
i.e., a period 5–15 years ahead, which can represent and cover the applicable development 65
cycles of construction robotics. Scenario approaches have successfully been used and 66
adapted to a variety of industries, such as automobile and consumer goods, where long-67
term projections are required for improved planning of fundamentals and decision-68
making on future product lines [9, 10]. In this study, the authors transfer and adapt the 69
basic scenario approach as a new tool in the field of construction robotics to structure the 70
development context as the fundamentals of systems engineering processes of 71
construction robots. Hong Kong is used as an initial test case setting for the study, 72
representing a locally confined and complex industrial ecosystem with a mature 73
construction industry [11]. 74
-
Accepted manuscript
4
The remainder of this paper is structured as follows. Section 2 outlines the background 75
of the study. Section 3 describes the integrated scenario approach developed for and 76
applied in this study. Section 4 introduces the scenario creation phase that develops the 77
scenarios from the prognosis and clustering of key influencing factors and then describes 78
them as snapshots of potential scenarios. Section 5 presents the scenario transfer phase 79
that further identifies contextual challenges and opportunities to generate strategic 80
planning for stakeholders under each scenario. Section 6 discusses the contributions and 81
limitations of the study. Section 7 concludes the paper and proposes future research 82
directions. 83
84
2 Background: concepts, advance beyond the state of the art, and the 85
test case 86
To set the stage for the methodological approach, this section introduces the relevant 87
terms and concepts in the context of construction robotics, analyzes the state-of-the-art 88
studies on the future of construction robots and forecasting techniques, and introduces the 89
initial test case setting (Hong Kong). 90
91
2.1 Key concepts in construction robot engineering 92
While construction automation and robotics generally cover a broad spectrum of 93
technologies, this study exclusively focuses on construction robots. Although a consensus 94
has not yet emerged on a clear definition of a construction robot [7], it often refers to 95
more sophisticated and intelligent types of machinery [12] with robotic features. To 96
facilitate the requirement analysis for systems design and development of construction 97
-
Accepted manuscript
5
robotics [12], a clear categorization is essential to thoroughly understand and differentiate 98
the development requirements and goals for construction robots [19, 23]. 99
Therefore, in this study, a two-dimensional perspective for the categories of robots 100
for buildings (Fig. 1) is proposed considering the building life-cycle and the level of task 101
integration. Usually, construction robots are, to some extent, task-specific (concrete 102
finishing, wall painting, bricklaying, facade cleaning, demolition, etc. [4, 13, 14]). In 103
addition, new forms of robots have emerged, such as adding and integrating aerial 104
approaches [15], exoskeletons for power augmentation [16], additive manufacturing 105
technologies [17], collaborative robots [18], and humanoid robot technology [19]. Such 106
robots help construction workers with more general tasks and can be considered to be 107
task-generic. Larger, complete systems are referred to as integrated automated and robotic 108
systems or automated/robotic on-site factories [20, 21]. 109
The development and systems engineering process for construction robots requires a 110
domain-specific, multidisciplinary, and phased approach, which was first conceptualized 111
by Hasegawa et al. [22] and Maeda [23], and then later advanced and developed by, for 112
example, Bock and Linner [24] and Linner et al. [25]. As systems engineering gains 113
prominence in other industries [26], the successful development of construction robots 114
over subsequent, linked, and iterative steps is dependent on a thorough understanding and 115
definition of the dynamic contexts surrounding such systems, as well as the “informed” 116
inference and management of associated requirements. 117
[insert Fig. 1 here] 118
119
-
Accepted manuscript
6
2.2 Previous systematic studies on the future construction robots and novelty of a 120
scenario approach 121
Studies on the future of construction robots and possible future developments have 122
focused primarily on existing barriers to implementation and future promises. 123
Warszawski and Navon [27] identified four fundamental reasons for the minimal success 124
of robots in building construction: insufficient development, unsuitable building design, 125
inadequate economic justification, and managerial barriers. In response, they proposed 126
strategies for more efficient future implementation. Mahbub [7] analyzed and ranked the 127
barriers to infiltrating construction automation and robotics in Japan, Australia, and 128
Malaysia with different levels of usage, and briefly speculated on the plausible future. 129
Bock [2] categorized the future trends in construction automation and robotics into five 130
areas: robot-oriented design, robotic industrialization, construction robots, site 131
automation, and ambient robotics. Quezada et al. [28] explored the future of the 132
construction workforce in Australia and discussed the impacts of construction automation 133
and robotics on future labor. Bogue [13] discussed and illustrated the current uses and 134
potential role of robots in improving the construction industry, with examples of different 135
classes of robots. These studies provide insightful understandings of the application 136
challenges or future prospects of construction robots, with focus on the details of 137
promising technologies. However, there are few studies that examine the future potential 138
of construction robots in a comprehensive and systematic manner. 139
Multiple sociotechnical issues could lead to uncertainties and eventually allow many 140
possible deployments for robotic construction. The possibility of potential scenarios to 141
reduce ambiguity and understand complexity in the design of construction robots is 142
therefore of great importance, but remain unexplored. In this regard, a scenario approach 143
-
Accepted manuscript
7
[9, 10] would be key to holistically understand and systematically develop scenarios, i.e., 144
behaviors and future states of the robotic ecosystem to which the requirements and 145
development activities for robots shall be subsequently oriented. Such approaches could 146
provide adequate information to support the system engineering processes of construction 147
robots. Construction robotics is not yet an established field and lacks know-how and 148
experience about basic scenarios and requirements. Thus, domain-specific forecasting 149
techniques would provide developmental and structural contexts as well as allow design 150
inputs. This study takes an initial step toward addressing this gap by providing, for the 151
first time, systematic scenarios to examine and structure the future robotic ecosystem 152
context of construction robots. Specifically, a scenario approach integrating qualitative 153
and quantitative methods was developed and applied to systematically develop the 154
possible narratives and solution spaces for the future of construction robots, using Hong 155
Kong as a test case. The scope of this study is limited to building work; consequently, 156
robots working in civil engineering projects are excluded. The scenario approach can 157
serve as the initial steps toward an effective tool to inform and guide construction robot 158
developments, reducing requirements uncertainty by establishing the structure and 159
formalizing the scenarios from which design inputs can be inferred for the later systems 160
engineering process. 161
162
2.3 Initial test case setting: Hong Kong construction industry 163
The building and construction industry is fundamental to urban growth and plays a 164
pivotal role in achieving the rapid economic development of Hong Kong [11]. The 165
industry is characterized by a small group of large local contractors, who make up a large 166
number of small-sized local construction companies that support the high level of 167
-
Accepted manuscript
8
subcontracting with the support of many overseas contractors [29]. Concerning the 168
building sector, work can be classified by building typologies into residential and 169
nonresidential, or public and private [30]. 170
Over the years, construction companies in Hong Kong have also developed their 171
expertise in qualified performance and gained a reputation in such a large and fast-172
growing market. However, along with these achievements, the industry faces challenges 173
on many fronts. The ever increasing demand for labor, accompanied by skills mismatch 174
(projected shortfall in skilled construction workers from 5,000 to 10,000 between 2019 175
and 2023) and aging labor force (37.7% of skilled and semi-skilled construction workers 176
were above the age of 55 in 2018) poses significant risks to the realization of a productive 177
industry [11, 31]. Moreover, the escalation of construction costs—Hong Kong had the 178
third-highest construction costs worldwide in 2018—and land shortage (predicted 179
shortfall of 1,200 hectares toward 2030) have made it difficult for construction companies 180
to make a profit [11, 32]. Site safety performance in terms of fatalities (22 industrial 181
fatalities in 2017) and accidents (3902 industrial accidents in 2017) remain unsatisfactory 182
[11]. The impact of construction work on the environment and general public in Hong 183
Kong is in a critical state in terms of carbon emissions, waste demolition, noise pollution, 184
and disturbance to surrounding areas [33]. All these issues are severely hindering the 185
sustainable development and continuous prosperity of the industry. 186
The industry has explored advanced technologies and intelligent approaches to tackle 187
the challenges and thrive in dynamic circumstances. The conception and implementation 188
of advanced construction technologies (e.g., [11, 34, 35]) have been facilitated by its key 189
actors throughout the last couple of years. For example, it is a pioneer in prefabrication 190
since the early 1970s, yielding significant economic and environmental benefits [36], and 191
-
Accepted manuscript
9
is currently promoting the adoption of modular integrated construction [34, 37] as a more 192
advanced method for off-site construction. Major local construction companies are 193
progressively investing in innovative technologies such as robotic arms, exoskeletons, 194
and 3D printing for construction, which hugely invigorate and inspire the remainder of 195
the industry [38]. Meanwhile, the government has provided continuous financial and non-196
financial support to the industry to create a more fertile environment for innovation and 197
technology R&D. The Construction Industry Council (CIC), as a statutory body, launched 198
the Construction Innovation and Technology Application Centre in 2017 to accelerate 199
information sharing and practices on the latest construction technologies [39]. 200
Furthermore, CIC also established a new institute in 2018 to cultivate higher caliber and 201
professional practitioners for the construction industry [1]. 202
In short, for this study, Hong Kong is considered a valuable test case. Firstly, the 203
significance of the built environment has always been highlighted by the government as 204
the overwhelming theme for future development [1, 40]. Secondly, Hong Kong as a high-205
rise, high-density metropolis is facing severe labor and cost challenges that are strongly 206
linked to unperformed productivity, which are also global problems especially in modern 207
cities [11]. These challenges create the ideal environment for utilizing construction 208
robotics in Hong Kong and similar cities or economies dealing with increasing density 209
and urbanization, a growing number of high-rise buildings, an aging workforce, and labor 210
shortage. Despite regional differences, the Hong Kong case can be considered as a global 211
reference for understanding the essential elements (e.g., conceptual, functional, 212
operational, and environmental requirements of system development) of construction 213
robots against different future motives and challenges. 214
215
-
Accepted manuscript
10
3 Methodology 216
Scenario approaches are widely applied to predict and understand the potential 217
outcomes of technological changes such as directions, rate, characteristics, and impacts, 218
incorporating the uncertainties of complex long-term development for investment and 219
policy strategizing [9]. The term “scenarios” has different definitions [10]. In this study, 220
a scenario is defined as a plausible combination of alternative developments in critical 221
dimensions [41]. The inherent benefit of the scenario approach is the consideration of a 222
range of possible future alternatives, thereby allowing stakeholders and practitioners to 223
have alternative views of the future to properly define the requirements and reduce the 224
risks of making the wrong decisions [42]. This is preferable for this study, which could 225
provide contextual awareness of the potential alternatives for construction robots in future 226
utilization, thereby supporting the strategic requirements for the definition and decision-227
making by different stakeholders. 228
There is no universal scenario method [43]. In order to provide rich and complicated 229
profiles to establish potential prognoses theoretically and systematically, scenario 230
approaches have been conducted differently based on the backgrounds, goals, and steps 231
applied [42]. Previous scholars [42, 44] have compared various scenario approaches and 232
summarized the fundamental steps to construct scenarios. Besides, since the scenario 233
approach heavily hinges on the accuracy of assumptions, several strategies of 234
improvements have been proposed by integrating other methods like mathematical 235
models [45], system dynamics [46], and fuzzy theory [47] to improve the scenario 236
formation and description. Nevertheless, little attention has been paid to improve the 237
identification of key influencing factors in the scenario approach, which is one 238
fundamental step to provide the critical components for the scenarios. Cross-impact 239
-
Accepted manuscript
11
analysis (CIA) [48] and complementary qualitative analysis [42] are standard methods 240
for the identification of key influencing factors. However, CIA considers only the direct 241
impacts of factors to capture a causal relationship analysis, and complementary 242
qualitative analysis is inherently subjective to ensure consistent results [42, 48]. A more 243
reliable identification of key influencing factors through causal relationship analysis 244
could be achieved with the consideration of both direct and indirect impacts of factors in 245
a quantitative manner. Consequently, this paper proposes using the decision-making trial 246
and evaluation laboratory (DEMATEL) method to identify key influencing factors by 247
analyzing interdependence between factors in a causal diagram and providing contextual 248
understanding in scenarios [49]. A modified fuzzy DEMATEL method (see in Appendix 249
A) is proposed, which integrates the fuzzy set theory into DEMATEL [50] with further 250
modification to address the ambiguous issues within human judgments and different 251
judgment criteria. 252
Based on previous studies [9, 42], an integrated scenario approach was applied as a 253
new tool to structure the context for future construction robot development. The approach 254
consists of three phases, with eight steps as follows (Fig. 2). 255
[insert Fig. 2 here] 256
• Step 1 - Define the object of analysis: The first step is to define the object of analysis 257
and scenario field. In this study, the focus is on the utilization of construction robots 258
for buildings in the context of Hong Kong within a time window of 18 years from the 259
baseline year (2017). The relevant definitions and background have been outlined in 260
the background section. 261
• Step 2 - Identify influencing areas and influencing factors (IFs): A multidimensional, 262
multilevel, sociotechnical conceptual framework was proposed based on the 263
-
Accepted manuscript
12
multilevel perspective (MLP) theory [51] and the PESTEL (political, economic, 264
sociocultural, technological, environmental, legal) model [6]. Drawing on the 265
framework, the scenario field was divided into influencing areas and the IFs in each 266
area were identified based on a holistic literature review and brainstorming. Semi-267
structured interviews with 20 experts were conducted to verify the identified IFs. 268
Purposeful sampling was used to ensure the sample covering different stakeholder 269
groups for representation. Fig. 3 illustrates the whole procedure in this step. 270
[insert Fig. 3 here] 271
• Step 3 - Identify key influencing factors (KIFs): KIFs can be determined by pairwise 272
influence analysis of the IFs. A modified fuzzy decision-making trial and evaluation 273
laboratory (DEMATEL) method was proposed and applied to analyze the causal 274
interrelationships between the IFs, and identify KIFs [52]. Data were collected 275
through a DEMATEL questionnaire survey (see Appendix B) of 18 professionals 276
using purposeful sampling. The identified KIFs were further validated and finalized 277
by two focus group meetings involving 16 professionals. The whole procedure in this 278
step is illustrated in Fig. 4. 279
[insert Fig. 4 here] 280
• Step 4 - Derive projections of KIFs: The purpose of deriving KIF projections is to 281
obtain the fundamental components for developing plausible scenarios. The 282
appropriate number of projections bears on the number of scenarios to be created [42]. 283
Typically, 3 to 5 scenarios are recommended by most researchers [43]. In this study, 284
a four-scenario outcome was envisaged, since a three-scenario framework could pose 285
the risk of centering on the middle scenario, and five scenarios may not be justifiable 286
[43]. Possible developments of each KIF were projected within the target time 287
-
Accepted manuscript
13
window based on documentation review and analysis, and then examined and 288
finalized through two focus group meetings. One meeting focused on technology 289
while the other on the application environment. The participants were asked to either 290
select or initiate new KIF projections based on the referential ones provided by the 291
research team. The past development and status quo of each KIF were briefly 292
introduced, with participants tasked with forecasting the most relevant factors 293
according to their professional backgrounds. Fig. 5 illustrates three examples of how 294
to develop the KIF projections. KIFs represented as red dots from IFs were projected, 295
respectively. 296
[insert Fig. 5 here] 297
• Step 5 - Cluster projections: The purpose of clustering projections is to create raw 298
scenarios from the combination of projections that are most likely to occur 299
simultaneously. The procedure in this step is outlined in Fig. 6. The consistency matrix 300
[9], with pairwise comparisons of consistency, was applied to cluster compatible 301
projections into bundles. In each case, all projections of a factor were first juxtaposed, 302
element by element, with all projections of the other factors. The consistency 303
assessment was based on a five-point scale with 5 = strong consistency (strong mutual 304
support) and 1 = strong inconsistency (complete opposition). Then, projection bundles 305
were obtained with all possible combinations of KIF projections (each bundle 306
containing one projection from each KIF), and a consistency value was calculated and 307
ranked for each bundle. The most consistent bundles can then be obtained based on 308
the consistency value with a defined quality level for calculation. Supported by the 309
ScMI AG software [53], similarity, referred to as “distance,” can be calculated 310
between any two bundles according to differences in projections, where one different 311
-
Accepted manuscript
14
KIF projection is calculated as “1.” Similar projection bundles (each bundle 312
containing one projection from each KIF) are clustered into groups based on 313
connectivity-based clustering (hierarchical clustering) to develop the raw scenarios. 314
[insert Fig. 6 here] 315
• Step 6 - Describe scenarios: The raw scenarios were further interpreted and described 316
pictorially to create the final scenarios, considering the trends of KIFs in the raw 317
scenarios and their impacts. Scenarios could then be described in terms of what and 318
how construction robots will be utilized for buildings in Hong Kong within the 319
targeted timeline, considering the sociotechnical configuration of the industry. The 320
developed scenarios were further verified by experienced and senior professionals 321
from building contractors through a questionnaire survey in terms of probabilities and 322
additional concerns, considering that contractors are direct adopters of construction 323
robots. 324
• Step 7 - Analyze disruptive events: Disruptive events, also known as “black swan” 325
events, are future events with low probabilities of occurrence but with high impact on 326
other events and the environments in which they occur. The possible disruptive events 327
were identified, and their effects on scenarios were discussed. 328
• Step 8 - Elaborate strategies: Implications of the scenarios were explored with regard 329
to driving forces, opportunities, and challenges under each scenario. Strategies for 330
different stakeholders under each scenario were also developed accordingly. 331
332
The study involves data collection through literature review and brainstorming by the 333
research team, as well as interviews: two questionnaire surveys and focus group meetings 334
-
Accepted manuscript
15
with professionals in different steps. Details on the professionals who participated in 335
interviews, surveys, and focus group meetings in relevant steps are presented in Table 1. 336
[insert Table 1 here] 337
Aligned with Fig. 2, the results and findings of scenario creation and scenario transfer 338
are presented in Sections 4 and 5. 339
340
4 Scenario Creation 341
This section introduces the Scenario Creation phase, where scenarios are created to 342
describe the possible future use of construction robots for building construction in Hong 343
Kong, according to the methodology presented in Section 3. 344
345
4.1 Identify influencing areas and influencing factors 346
As described in Section 3, a holistic literature review, brainstorming, and expert 347
interviews were undertaken to identify influencing areas and IFs based on the conceptual 348
framework developed from MLP theory [51] and PESTEL model [6]. The MLP has 349
emerged to explain and analyze socio-technical transitions as interactive processes of 350
changes in niches, regimes and landscape levels [51] and PESTEL is a useful analytical 351
tool to examine socio-technical factors in a multidimensional way [6]. The framework 352
could, therefore, allow a multidimensional multilevel analysis of a broad range of factors 353
affecting the successful transition of construction robots into the industry, as niche 354
developments in their infancy. Drawing on the framework, a critical literature review was 355
conducted on relevant studies (e.g., [2-4, 7, 12-14, 23, 25, 26, 54, 55]), through which 65 356
IFs were initially identified and later reduced to 25 IFs (after combination and refinement, 357
see Appendix B) according to economic, environmental, industry, political, sociocultural, 358
-
Accepted manuscript
16
and technological aspects [52]. The identified IFs were verified by relevant experts as 359
inclusive and influential to the future utilization of construction robots for buildings in 360
Hong Kong. 361
362
4.2 Identify key influencing factors 363
Through the aforementioned process of identifying KIFs, KIFs can be determined 364
through pairwise influence analysis of IFs using the modified fuzzy DEMATEL method. 365
The 25 IFs were assigned to formulate the matrix-based DEMATEL questionnaire survey 366
(see Appendix B) for KIF analysis and verified using focus group meetings [52]. Table 2 367
outlines the identified 11 KIFs, together with their descriptors for making projections. 368
[insert Table 2 here] 369
370
4.3 Derive projections of key influencing factors 371
To provide evidence for projection development, the past development and status quo 372
of each KIF were investigated through the review of relevant literature and documents 373
[32, 56-58]. By considering government planning and policies [40, 56], up to four 374
possible developments for each KIF were projected in the target time horizon by the 375
research team as the reference point. The proposed projections were examined and 376
finalized through two focus group meetings. After integrating perceptions from 377
professionals, a full list of KIF projections was formulated, as shown in Table 3. Some 378
factors have been described as having only one clear projection, such as "demand for 379
green buildings," with consensus on an increased state. 380
[insert Table 3 here] 381
382
-
Accepted manuscript
17
4.4 Cluster projections 383
All plausible combinations of the future were constructed based on consistency 384
analysis, while the most consistent projection bundles were obtained according to the 385
consistency values, which were clustered into groups based on the similarity of 386
developing the raw scenarios. The results of raw scenarios are presented in Fig. 7. Each 387
circle represents a projection bundle, and the similarity of 97 projection bundles defines 388
the spatial location. All projections in a bundle in one cluster were combined and 389
alternative projections were identified in raw scenarios based on the calculated 390
occurrence proportion p in the scenario [9]. The alternative projections were validated in 391
the next step according to their fitness to scenario descriptions. 392
[insert Fig. 7 here] 393
394
4.5 Describe scenarios 395
Whilst it is impossible to accurately predict how construction robots will be utilized 396
in the future, the scenario narratives indicate areas in which a reasonably plausible future 397
may be illustrated through certain influencing conditions. Therefore, for each scenario, a 398
sense of the context for technology utilization is offered, considering technological 399
features, the depth and breadth of usage, and essential characteristics of the industry. 400
Each raw scenario as a projection cluster in Fig. 7 is given a characteristic name. Detailed 401
descriptions are provided below. 402
403
4.5.1 Scenario 1: Bottleneck 404
In this scenario, the industry will experience few disruptive changes with construction 405
robots, which have not been well developed and applied as expected owing to the multiple 406
-
Accepted manuscript
18
“bottlenecks” of economic justification, technology usability and availability, and 407
industry culture and structure. In short, they are not appealing enough to the construction 408
industry in Hong Kong. Traditional trades will remain in high demand. Although there is 409
increasing use of ICT and prefabrication, the industry will continue suffering from the 410
issues of an aging workforce, scarcity of skilled labor, and high construction costs. 411
Despite the rising awareness in sustainability and increasing demand for green buildings, 412
primary attention will be paid to the operation stage, and environmental issues during the 413
construction stage will remain. Construction productivity may even decrease, as more 414
man-hours may be needed for completing the same work, compared with the status quo, 415
owing to the increasing geographical difficulties, as well as increased quality and 416
functional requirements. The government may tighten the financial support for 417
construction robotics and focus more on other innovative technologies for construction. 418
419
4.5.2 Scenario 2: Age of Iron Worker 420
This scenario outlines a continuously improving industry centered on robot-assisted 421
construction workers, like “Iron Men.” BIM and many other ICT tools will be 422
substantially used to facilitate the digital transformation of the industry, laying the 423
groundwork for implementing robotics. Construction robotics will begin to emerge in 424
some applications to support human workers, facilitated by technological advancement 425
and increasing labor challenges. Most likely, the industry will utilize well-developed, 426
general-purpose robots, like exoskeletons and drones, to assist the aging workforce and 427
manage the labor shortage. Furthermore, heavy machinery and equipment will be 428
developed toward highly integrated, automated, and intelligent levels. Risk aversion 429
across the sector will continuously shape a conservative culture, such that the industry 430
-
Accepted manuscript
19
practitioners will remain reluctant to use innovative technologies that radically change 431
conventional practice, like humanoid or autonomous robots, highly integrated on-site 432
robotics factory. 433
434
4.5.3 Scenario 3: Dynamic Co-evolution of Robotization and Modularization 435
In this scenario, the use of construction robots is strongly linked to increasing 436
modularization and prefabrication, which will provide a substantial boost to construction 437
productivity and sustainability, as well as ease the skilled labor shortage. The uptake of 438
modularization or increasing rate of prefabrication could somewhat restrain the demand 439
for extensive use of on-site robots; however, these also create a more controllable and 440
favorable environment which favors the use of certain kinds of robots or systems to assist 441
or perform work such as site management, logistics, material handling, and on-site 442
assembly. The government as the main client broadly encourages off-site construction in 443
government procured projects, e.g., for achieving greater efficiency in delivering massive 444
housing programs, while clients in the private sector become the leading force for 445
automation and robotics in particular projects. A fundamental change may exist in 446
traditional patterns for building construction with entirely different on-site needs for 447
human labor, but very likely, the future will witness a “dynamic co-evolution of 448
robotization and modularization.” 449
450
4.5.4 Scenario 4: Rise of the Robots 451
This scenario represents a future in which the Hong Kong construction industry is 452
aggressively pursuing robotic construction technologies to deal with sociotechnical 453
challenges, thus contributing to a productive, sustainable, and knowledge-intensive sector. 454
-
Accepted manuscript
20
However, this “rise of the robots” could generate disruptive changes. Technologically, 455
construction robots will be maturing at an impressive rate with increased usability, 456
underpinned by advancements in sensor technology and artificial intelligence (AI). Hong 457
Kong, as a vibrant adopter, will benefit from its positive socio-technical environment for 458
construction robots to attract international investors and technology developers, while its 459
intrinsic technological capability will also be boosted with increased R&D activities. 460
Owing to the high availability of technology, along with the reduced capital cost and 461
payback period, a diverse array of on-site robots or even those in the form of a highly 462
integrated robotic site will be utilized for building construction and deconstruction in 463
Hong Kong. The government will have provided sufficient financial support and initiated 464
incentive schemes in robotics research and applications in construction. Specific guidance 465
and standards for construction robots will have also been developed. The industry will 466
undergo major changes in the allocation of labor and work profiles. 467
468
4.5.5 Perspectives on future scenarios by building contractors 469
The probabilities of the four developed scenarios were further assessed by 94 470
professionals from different building contractors in Hong Kong using a questionnaire 471
survey. Respondents were asked to assess the probability of each scenario regarding the 472
use of construction robots in 2035 in Hong Kong, described by critical trends, using a 473
five-point Likert scale, and to state their additional concerns in their response to an open-474
ended question. As seen in Fig. 8, the four scenarios were generally agreed as probable 475
(mean>3), with higher probability of scenario 2 (Age of Iron Worker) and scenario 3 476
(Dynamic co-evolution of Robotization and Modularization). The responses to the open-477
ended question are summarized as follows. 478
-
Accepted manuscript
21
• The future utilization of construction robots may not change if the “no change no fault 479
culture” remains across the sector, which leads to fear of failure by using construction 480
robots. 481
• There should be more support from developers and the government to make the future 482
utilization of construction robots more feasible. 483
• Construction robots could be used to replace low-level, high-risk jobs in the future. 484
• It is unknown whether intelligent robots can replace a large number of workers in the 485
future. Nevertheless, if so, it will result in unemployment for many workers and 486
impose additional burdens on the government. 487
[insert Fig. 8 here] 488
489
5 Scenario Transfer 490
The developed scenarios have revealed that the future alternatives of construction 491
robots for building in Hong Kong are strongly coupled with the sociotechnical 492
environment and developments. This section covers the Scenario Transfer phase, in which 493
disruptive events are further analyzed, and strategies to enable scenario transfer and move 494
the industry forward are elaborated, with references to relevant studies or cases. 495
496
5.1 Analyze disruptive events 497
Possible disruptive events that could interfere with the future development of 498
construction robots are identified through a historical review of the development of 499
construction robotics. Starting from the late 1970s, the Japanese have surpassed the 500
progress of other regions of the world in the area of construction robots with extensive 501
R&D activities. However, its economic crisis in the 1990s sharply slowed down 502
-
Accepted manuscript
22
technological advancement and restrained robotic applications [7]. In this context, 503
economic turmoil associated with social instability is a possible disruptive event, which 504
may force the implementation of construction robots toward Scenario 1, and strongly 505
deaccelerate or disrupt Scenarios 3 and 4. Strengthening collaboration and 506
communication on R&D for robots can help to reduce the negative impacts of such an 507
event. Another possible disruptive activity interfering with the use of construction robots 508
would be the softened regulatory control on cheap foreign construction workers that 509
diminish the allure of robots to solve labor issues. All scenarios will be affected, but rapid 510
technological progress to improve the economic effectiveness of robotics will erode 511
whatever cost advantage the imported labor enjoys. 512
513
5.2 Elaborate strategies 514
5.2.1 Identification of driving forces, opportunities, and challenges 515
The developed scenarios highlight how the utilization of construction robots may 516
change in Hong Kong in the coming decades, where opportunities and challenges co-exist. 517
The driving forces of each scenario should first be identified to determine appropriate 518
strategies that maximize the opportunities and minimize potential risks. Considering their 519
influence on other factors and their distinctiveness in different scenarios, driving forces 520
could be represented by KIFs. These KIFs were found to be extremely influential in the 521
factor analysis in Section 4.2 and are the main causes of variations among scenarios. 522
Seven KIFs were defined and further classified as primary (i.e., initial investment cost 523
and economic performance associated with robots, ease of use of robots, availability of 524
robotic technology, and culture of innovation in the industry) and secondary (i.e., the 525
scale of prefabrication, government support of robotics applications in construction, and 526
-
Accepted manuscript
23
work structure and organization) based on whether they are “uncontrollable” or 527
“controllable” by regulators and industry [9]. Opportunities and challenges can be 528
identified from the corresponding outcomes of those significantly affected KIFs and other 529
IFs under the scenarios. 530
Table 4 provides a detailed scenario-based description of driving forces, and the 531
identified opportunities and challenges facing the industry. Technologically, Scenario 1 532
offers chances for robotics startups to be the technology pioneers and leaders, along with 533
high risks of being frustrated by indifferent market; Scenario 2 and 3 provide golden 534
opportunities for construction robotics, under the “survival of the fittest” pressure from 535
the industry practices; Scenario 4 creates the stage for disruptive construction robotics to 536
takeoff, albeit challenged by a highly competitive market. 537
[insert Table 4 here] 538
539
5.2.2 Recommended strategies for stakeholders 540
Based on Table 4, policies can be proposed for all stakeholders along the value chain, 541
including industry (also conventional construction companies and emerging companies 542
for construction robots), government (its departments and agencies), and academia in 543
Hong Kong. The analysis should work backward from the implications of the four 544
scenarios and their impacts on the current situation. For industry practitioners, they should 545
fully seize technological opportunities by integrating different scenarios into their 546
technology strategy and business plans. For government and academia, strategic 547
recommendations should focus on how to realize desirable scenarios for a flourishing 548
future and to mitigate undesirable outcomes. Primary driving forces imply the 549
technological uncertainties can alter utilization behavior, which are uncontrollable, but 550
-
Accepted manuscript
24
trends can be learned by decision-makers and the government to make correct scenario 551
prognoses and corresponding strategic plans. Accordingly, strategic repertoires for each 552
scenario are explored where the undergoing trajectory is most likely to arrive. 553
Scenario 1 (Bottleneck) is, in general, an unexpected scenario. Although substantial 554
employment opportunities are provided in this scenario, the industry is slowly moving 555
forward. The escalation of construction costs will continuously lead to increasing 556
property prices and rentals, causing severe livelihood and other social problems. 557
Therefore, the following strategies and measures are proposed to prevent the occurrence 558
of the scenario or change its development trajectory. 559
• Upgrading existing machinery and equipment. The conservative culture of the 560
industry could inhibit robotics development and maintain the dominance of traditional 561
trades. To overcome severe labor challenges and boost productivity, the gradual 562
upgrading of existing machinery and equipment toward automation and intelligence 563
[e.g., 59] should be fostered in the whole industry. 564
• Emphasizing technology usability and interoperability. Robotics startups should work 565
closely with the industry and develop real-world cases with proven benefits to raise 566
industry awareness and improve technology usability. A later step should be the focus 567
on robot-oriented design [24] to guide and enable the applicability, simplicity and 568
efficiency of robots in real-world practice. 569
• Providing a fertile environment for construction innovation. The government should 570
create a productive environment for construction innovation by providing funding 571
support and incentives to construction companies and academia for developing and 572
deploying promising technological innovations, including construction robotics. This 573
could further foster the culture of innovation in the industry. 574
-
Accepted manuscript
25
• Strengthening talent cultivation and labor management. The government and 575
academia should provide sufficient education and training on the latest innovations in 576
the construction industry. Traditional contractors, especially specialized 577
subcontractors, should put more effort into labor training and labor resource 578
management. The entire industry should work together toward safer and healthier 579
construction sites to attract the young generation. 580
Scenario 2 (Age of Iron Worker) is a transitional stage for the industry toward the 581
direction of a better-equipped workforce. The technological, economic and cultural issues, 582
to a certain extent, restrict the utilization of more disruptive construction robots in Hong 583
Kong. Problem-solving, combined with the rational use of robots, should focus on the 584
following strategies. 585
• Promoting diversified innovations. The government should capitalize on 586
advancements in diversified innovations, including robots and other emerging 587
techniques, and vigorously promote their real-world adoption to tackle challenges of 588
stagnant productivity, high construction cost and housing affordability. Accordingly, 589
the partnership between the industry and universities should be encouraged, financial 590
and non-financial support should be provided, and enhanced education and pertinent 591
training programs should be established. 592
• Grasping technological opportunities of construction robots. Construction companies 593
should explore investment opportunities in construction robotics in the global arena 594
and evaluate potential technologies within the local market to allocate their R&D 595
resources and efforts accordingly. The emphasis should be on robot-oriented design 596
and development [24] to promote the application of construction robots. It is 597
recommended for the industry to first focus on more affordable wearable robotics or 598
-
Accepted manuscript
26
assistant robotics that are well developed in other industries (e.g., [15, 16]), and pay 599
attention to those trades facing severe labor problems. 600
• Promoting inter- and intra-industry collaborations. The industry, as a whole, should 601
consolidate their inter- and intra-industry communication and cooperation on robotic 602
application-oriented research, technology transfer, and real-world trials and develop 603
new business models under the sharing economy [28] to share the benefits and risks 604
associated with robotic implementation. 605
• Alleviating labor shortage via a multipronged approach. The government should 606
adopt a multipronged approach to solve aging labor and shortage issues [1] regarding 607
supporting technological research on labor-saving, promoting education and training, 608
improving welfare policy for construction workers, strengthening recruitment 609
informatization, etc. 610
Compared to Scenarios 1 and 2, Scenario 3 (Dynamic Co-evolution of Robotization 611
and Modularization) demonstrates an innovation-based industry with diversified 612
advancements, led by robotization and modularization. Strategic plans should be 613
formulated to promote technological mutualism or symbiosis [37] to maximize the 614
benefits, as well as address potential labor surplus caused by the large-scale use of off-615
site construction and on-site automation. 616
• Grasping multiple technological opportunities. Construction companies should 617
evaluate the opportunities under new construction methods and technologies, and 618
upgrade their practices, operational processes, or business models accordingly. 619
Companies for construction robots should seek to explore business opportunities 620
regarding the interoperability and compatibility of robots in construction sites via 621
increasing prefabrication [36] and modularization [37]. 622
-
Accepted manuscript
27
• Transmitting project-based knowledge and experience. Project-based knowledge and 623
experience regarding different robots and modularization (or increasing 624
prefabrication) should be learned and transferred industrial-wide. The government 625
and academia should investigate the merits and drawbacks of technology applications 626
within the context of projects, and establish technical guidance for the industry to take 627
optimized technological solutions. 628
• Promoting technological mutualism and integration. More cross-disciplinary 629
research should be encouraged toward robotics and modularization as well as other 630
innovations through specialized R&D institutions, to facilitate mutualism and 631
integration of robotics and off-site technologies, and to promote benign competition 632
[37]. The government should improve its approval process for building projects to 633
reduce the institutional barriers for adopting innovative construction technologies. 634
• Eliminating labor surplus issues. The government and academia should set up 635
effective training and skills programs to push construction workers into new skills 636
and career areas of modular construction and robotics training. 637
Scenario 4 (Rise of the Robots) seems to be the most promising configuration to 638
unleash the power of construction robots for a productive industry. Robotic technologies 639
are utilized vastly more effective and fruitful through advances in sensor technology, 640
materials, and AI, forming a fiercely competitive market. Meanwhile, the rise of robots 641
could significantly reduce labor demand and potentially result in a high industry 642
unemployment rate. The following strategies are thereby proposed. 643
• Strengthening the competitiveness of construction companies. Construction 644
companies should keep pace with worldwide technological innovations for 645
construction robots to avoid being technological laggards and losing competitiveness 646
-
Accepted manuscript
28
in the market. Small construction companies should emphasize the development of 647
originality and creativity, either in technology or soft aspects. 648
• Emphasizing the wider benefits of technology. The government should promote the 649
use of robots for a sustainable transition and formulate standards to evaluate the 650
sustainability of robotic technologies in projects [12, 17, 60]. 651
• Eliminating labor surplus issues. The government and academia should establish 652
training programs to advance the skillsets of construction workers, develop new 653
career areas emerging from construction robotics, and continuously nurture traditional 654
talents for irreplaceable construction tasks to fit the new industry. The government 655
should also provide adequate support to encourage career changes of traditional 656
construction workers to other sectors. 657
Although the focus of strategies is different for each scenario, the crucial role of 658
workers is highlighted in the future utilization of construction robots. Construction robots 659
are expected to solve severe labor issues, while also generating physical, ethical, legal 660
and social issues in human-robot collaboration [17]. Besides, the transition and 661
development of staff to be more open-minded and knowledgeable [28] is essential for 662
strategic planning of all scenarios. 663
664
6 Discussions 665
In the present study, the authors have, for the first time, applied a systematic scenario 666
projection approach in the field of construction robotics. Hence, they made an initial step 667
toward the application of a key tool for the future comprehensive development and 668
systems engineering of construction robots, by enabling industries to structure the 669
solution space, and thus mitigate uncertainties in the requirements engineering and 670
-
Accepted manuscript
29
management process. The development of construction robots should start now in 671
preparation of a largely unknown future. Starting with the use of the Hong Kong 672
construction industry as an initial test case, the findings of our study are generalized 673
toward and classified into three categories: theoretical dimension, methodological 674
dimension, and practical dimension. 675
• Theoretical dimension: Applying the multidimensional, multilevel systems 676
perspective, this study makes a theoretical contribution to the field of construction 677
robots by comprehensively illustrating how utilization can be shaped in sociotechnical 678
contexts. Dubois and Gadde [61] argued that innovation often suffers in the 679
construction industry as a loosely coupled system. For construction robots, the 680
dialectics [30] of couplings should be emphasized in terms of technology-oriented 681
cooperation patterns and other technology alternatives. 682
• Methodological dimension: The study demonstrates and verifies the scenario 683
approach as a scholarly research methodology [10], as well as the need for a 684
combination of interdisciplinary emphasis and scenario exploration methods [42]. In 685
this study, the fuzzy theory and DEMATEL method were integrated in the KIF 686
exploration step in the scenario approach, which addresses the ambiguous issues in 687
judgment-based factor analysis, and enables a contextual understanding of the 688
scenarios. The scenario approach serves as the initial steps toward a novel tool in the 689
context of systems engineering for the future requirements definition and 690
specification of construction robots. The whole methodology can also be applied in 691
other foreseen studies for technology development and utilization. 692
• Practical dimension: Practically, this study tested the approach with the Hong Kong 693
case to understand the future of construction robots and their impacts on the building 694
-
Accepted manuscript
30
industry (solution space) from evidence-based analysis. The findings enable 695
stakeholders to fully seize the technological opportunities and systematically shed 696
light on the potential risks, uncertainties, and design factors and inputs. The study 697
supplements existing studies [7, 13, 28] by offering a multiplicity of possible future 698
scenarios and targeting the specific technological application of construction robots. 699
In particular, policy and decisions have been proposed for the adjustment of the 700
identified “controllable” driving forces to modify or adapt the development of the 701
identified “uncontrollable” driving forces, thus steering the utilization and 702
development of construction robots toward the desired direction. In this respect, 703
technical features and industrial culture are key to the future utilization of construction 704
robots, whilst the government could exert a multifaceted influence to foster favorable 705
utilization. More specifically, the findings also echo previous literature in 706
emphasizing the importance of technology interaction/compatibility explorations [35] 707
and worker-oriented studies [15, 28] of construction robots. The application of the 708
developed scenario approach to Hong Kong as a representative test case has 709
demonstrated that this new tool can deepen the understanding of and structure the 710
dynamic contexts (robot ecosystem) around potential technological development 711
targets. Regarding the applicability of the tool for other economies, the differences in 712
KIFs and their projections could result in different scenarios and associated strategies. 713
714
Despite the theoretical, methodological, and practical contributions, limitations were 715
also identified in this study. Firstly, the combination and reduction of initially identified 716
IFs might affect the integrity and correlation of factors, which could lead to the neglect 717
of some essential elements in the following analyses. Secondly, consistency evaluation 718
-
Accepted manuscript
31
of projections was conducted by the scenario team, which is inevitably subjective and 719
could influence the outcome of the scenarios. Thirdly, the accuracy of assumptions made 720
during the whole approach is not sufficiently assured, although the integration of fuzzy 721
logic theory improved the reliability of judgments in KIF identification. Hence, future 722
improvement of the integrated scenario approach for more robust technological 723
forecasting is expected. 724
725
7 Conclusions 726
The lack of systematic approaches to explore future scenarios and solution spaces has 727
become a major challenge in the conception and systems engineering process in 728
construction robot development. Therefore, the research presented in this paper attempts 729
to take the first step toward the development of a new tool for the construction robotics 730
field that can inform and support actual requirements management and systems 731
engineering processes by scenario-based analysis. The Hong Kong construction industry 732
was used as an initial test case. In this context, plausible scenarios for the utilization of 733
construction robots for buildings (in Hong Kong) for the year 2035, i.e., “Bottleneck,” 734
“Age of Iron Worker,” “Dynamic co-evolution of Robotization and Modularization,” and 735
“Rise of the Robots,” were systematically generated based on the integrated scenario 736
approach while considering the input and perspectives of professionals from both 737
academia and industry. The findings demonstrate that the future use of construction 738
robots is strongly associated with technological and social developments within the 739
economic and political ecosystem in the industry. “Uncontrollable” and “controllable” 740
driving forces, as well as opportunities and challenges under each scenario, were further 741
identified. Strategies for different stakeholders were explored to adjust the “controllable” 742
-
Accepted manuscript
32
forces to modify or adapt the development of “uncontrollable” forces affecting the 743
utilization of construction robots, in an attempt to maximize the potential opportunities 744
and mitigate the risks for the future prosperity of the industry. The crucial role of workers 745
is highlighted in managing and strategizing the future use of construction robots for all 746
scenarios. The study also has implications for the technical potential of different types of 747
construction robots; hence, it can serve as a valuable guide for the construction industry's 748
comprehensive development roadmap and prospects. 749
Future research will extend and further validate the integrated scenario approach 750
applied in this study as a novel tool for structuring the context of construction robot 751
development as part of the fundamentals of a systems engineering process for 752
construction robots. The findings presented in this paper can be expanded and integrated 753
into technology road-mapping techniques (e.g., [37]) and comprehensive systems 754
engineering models (e.g., [26]) to conduct a fully integrated engineering effort [26] for 755
construction robot development. 756
757
Acknowledgments 758
This work was supported by a grant from the Germany / Hong Kong Joint Research 759
Scheme sponsored by the Research Grants Council of Hong Kong [Reference No. G-760
HKU704/15] and the German Academic Exchange Service [DAAD Grant No. 57217359]. 761
762
Appendix A. The modified fuzzy DEMATEL method 763
Considering the paper length, the detailed numerical analysis based on the modified 764
fuzzy DEMATEL method is not covered. The main steps in the modified fuzzy 765
DEMATEL method for KIF identification are as follows. 766
-
Accepted manuscript
33
1. Transferring collected data into positive triangular fuzzy numbers. Given the n 767
factors F={F1, F2,…,Fn}, K professionals are asked to evaluate the pair-wise influence 768
with a 4-point scale from [0, 1, 2, 3], representing the linguistic terms [No influence, Low 769
influence, Medium influence, High influence]. For each professional, an n x n initial 770
influence matrix can be generated as [ ]k
k ij n nX x = , where k is the number of professionals 771
with 1 k K . The collected influence score k
ijx represents the judgement of the 772
influence of factor i on factor j. The fuzzy logic is then introduced to deal with the 773
ambiguities of k
ijx . According to Chen and Hwang [62], k
ijx is transferred and expressed 774
in positive triangular fuzzy numbers ( ), ,k k k kij ij ij ija l m r= based on Table A.1. 775
[insert Table A.1 here] 776
777
2. Defuzzifying fuzzy numbers to crisp scores. The defuzzification step transfers the 778
fuzzy numbers of ( ), ,k k k kij ij ij ija l m r= back to the crisp scores kija , which can be performed 779
according to Converting Fuzzy data into Crisp Scores (CFCS) method [63] as follows. 780
Firstly, the fuzzy numbers of kija are standardized based on results from all professionals. 781
1 11( min ) / (max min )k k k k kij ij ij ij ij
k K k Kk Kl l l r l
= − − (1)
782
1 11( min ) / (max min )k k k k kij ij ij ij ij
k K k Kk Km m l r l
= − − (2)
783
1 11( min ) / (max min )k k k k kij ij ij ij ij
k K k Kk Kr r l r l
= − − (3)
784
Secondly, the left score (ls) and the right score (rs) can be calculated as: 785
/ (1 )k k k kij ij ij ijls m m l = + − (4)
786
/ (1 )k k k kij ij ij ijrs r r m = + − (5)
787
-
Accepted manuscript
34
Thirdly, the total normalized value nx can be computed as: 788
[ (1 ) ] / (1 )k k k k k k kij ij ij ij ij ij ijnx ls ls rs rs ls rs= − + − + (6)
789
Lastly, the crisp score kija of the transferred fuzzy assessment
k
ija can be computed as: 790
1 11min max mink k k k kij ij ij ij ij
k K k Kk Ka l nx r l
= + −( ) (7) 791
3. Normalizing and generating the average matrix. Based on the above defuzzification, 792
the new initial influence matrix of the professional k is obtained as [ ]kk ij n nA a = . Here, the 793
added normalization step is applied to obtain the normalized initial influence matrix 794
[ ]kk ij n nD d = , which is the mapping from k
ijd to [0, 1]. The commonly used method [49, 795
50] is adopted for the normalization as follows. 796
k k kD s A= (8) 797
where 798
11
1
maxk n k
ijji n
sa
=
=
(9) 799
Then, the average matrix A can be obtained, with the element (1 , )ija i j n calculated 800
as: 801
1
1 K kij ij
k
a dK =
= (10) 802
4. Calculating the normalized direct and indirect influence matrix. The average 803
matrix A can be normalized via equations (10) and (11) to calculate the normalized direct 804
influence matrix D. Similar to obtaining the transition matrix of a Markov chain, the 805
normalized indirect influence matrix ID can be computed from the normalized direct 806
influence matrix D. 807
-
Accepted manuscript
35
( )12 3 2
2
1h
h
ID D D D D D D
−
=
= + + + = = − (11) 808
where I denotes the identity matrix. 809
5. Obtaining the total influence matrix. The total influence matrix T containing both 810
direct and indirect influence can be then acquired based on the summation of D and ID 811
as: 812
( ) ( ) ( )1 121T D ID D D D I D D I D− −
= + = − + − = − (12) 813
6. Depicting the causal diagram. Suppose is the (i, j) element of total influence matrix 814
T, then the sum of the ith row ri (total influences of factor Fi on others) and the sum of 815
the jth column cj, (total influences of others on factor Fi) can be calculated as: 816
1
n
i ij
j
r t=
= (13) 817
1
n
j ij
i
c t=
= (14) 818
The importance degree r+c and net effect degree r-c can be computed. For factor Fi, ri+ci 819
is an index of the power of the influences per factor (a measure of the importance of the 820
factor), and ri-ci is an index of whether the factor has more impact on others or can be 821
impacted by others (a measure of the net effect). The values of r-c also categorize factors 822
into cause and effect groups [50]. When the value of r-c is positive, the factor belongs to 823
the cause group. Otherwise, it belongs to the effect group. The causal diagram [49] can 824
then be obtained by mapping the dataset of (r+c, r-c), which visualizes the complex 825
causal relationships among factors. Key factors can thereby be identified regarding the 826
importance and causality of the influence. 827
828
ijt
-
Accepted manuscript
36
Appendix B. DEMATEL questionnaire in KIF identification 829
[insert Fig. B.1 here] 830
References 831
[1] Chief Executive, The 2018 Policy Address: striving ahead, rekindling hope, Hong 832
Kong, 2018, https://www.policyaddress.gov.hk/2018/eng/pdf/PA2018.pdf, Accessed 833
date: 20 January 2019. 834
[2] T. Bock, The future of construction automation: Technological disruption and the 835
upcoming ubiquity of robotics, Automation in Construction 59 (2015) 113-121. 836
https://doi.org/10.1016/j.autcon.2015.07.022. 837
[3] E. K. Zavadskas. Automation and robotics in construction: International research and 838
achievements, Automation in Construction 19(3) (2010) 286-290. 839
https://doi.org/10.1016/j.autcon.2009.12.011. 840
[4] T. Bock and T. Linner, Construction Robots: Elementary Technologies and Single-841
Task Construction Robots, Cambridge University Press, New York, 2016. ISBN: 842
9781107075993. 843
[5] S. Kale and A. David, Innovation diffusion modeling in the construction industry. 844
Journal of Construction Engineering and Management 136(3) (2009) 329-340. 845
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000726. 846
[6] W. Pan, L. Chen, W. Zhan, PESTEL analysis of construction productivity 847
enhancement strategies: A case study of three economies, Journal of Management in 848
Engineering 35(1) (2018) 05018013. https://doi.org/10.1061/(ASCE)ME.1943-849
5479.0000662. 850
[7] R. Mahbub, An investigation into the barriers to the implementation of automation 851
and robotics technologies in the construction industry, Ph.D. thesis, Faculty of Built 852
https://www.policyaddress.gov.hk/2018/eng/pdf/PA2018.pdfhttps://doi.org/10.1016/j.autcon.2015.07.022
-
Accepted manuscript
37
Environment and Engineering, Queensland University of Technology, 2008, 853
https://eprints.qut.edu.au/26377, Accessed date: 5 January 2017. 854
[8] W. Pan, R. Hu, T. Linner, T. Bock, A methodological approach to implement on-site 855
construction robotics and automation: a case of Hong Kong, in: Proceedings of the 856
35th International Symposium on Automation and Robotics in Construction, 2018. 857
https://doi.org/10.22260/ISARC2018/0051. 858
[9] J. Gausemeier, A. Fink, O. Schlake, Scenario management: An approach to develop 859
future potentials, Technological Forecasting and Social Change 59(2) (1998) 111-130. 860
https://doi.org/10.1016/S0040-1625(97)00166-2. 861
[10] R. Ramirez, M. Mukherjee, S. Vezzoli, A.M. Kramer, Scenarios as a scholarly 862
methodology to produce “interesting research”, Futures, 71 (2015). 863
https://doi.org/10.1016/j.futures.2017.09.006. 864
[11] Development Bureau, Construction 2.0 - Time to change, Hong Kong, 2018, 865
https://www.hkc2.hk/en, Accessed date: December 2018. 866
[12] M. Pan, T. Linner, W. Pan, H. Cheng, T. Bock, A framework of indicators for 867
assessing construction automation and robotics in the sustainability context, Journal 868
of Cleaner Production 182 (2018) 82-95. 869
https://doi.org/10.1016/j.jclepro.2018.02.053. 870
[13] R. Bogue, What are the prospects for robots in the construction industry? Industrial 871
Robot: An International Journal 45(1) (2018) 1-6. https://doi.org/10.1108/IR-11-872
2017-0194. 873
[14] L. Cousineau and M. Nobuyasu, Construction robots: the search for new building 874
technology in Japan, ASCE press, Reston, USA, 1998. ISBN-10: 0784403171. 875
https://eprints.qut.edu.au/26377https://doi.org/10.1016/j.jclepro.2018.02.053
-
Accepted manuscript
38
[15] T. Rakha and A. Gorodetsky, Review of Unmanned Aerial System (UAS) 876
applications in the built environment: Towards automated building inspection 877
procedures using drones, Automation in Construction 93 (2018) 252-264. 878
https://doi.org/10.1016/j.autcon.2018.05.002. 879
[16] T. Linner, M. Pan, W. Pan, M. Taghavi, W. Pan, T. Bock, Identification of Usage 880
Scenarios for Robotic Exoskeletons in the Context of the Hong Kong Construction 881
Industry, in: Proceedings of the 35th International Symposium on Automation and 882
Robotics in Construction, 2018, pp. 40-47. http://doi.org/10.22260/isarc2018/0006. 883
[17] S. H. Ghaffar, J. Corker, M. Fan, Additive manufacturing technology and its 884
implementation in construction as an eco-innovative solution, Automation in 885
Construction 93 (2018) 1-11. https://doi.org/10.1016/j.autcon.2018.05.005. 886
[18] K. Afsari, S. Gupta, M. Afkhamiaghda, Z. Lu, Applications of Collaborative 887
Industrial Robots in Building Construction, in: 54th ASC Annual International 888
Conference Proceedings, 2018. 889
http://ascpro0.ascweb.org/archives/cd/2018/paper/CPRT178002018.pdf, Accessed 890
date: 1 March 2019. 891
[19] Advanced Industrial Science and Technology, Development of a Humanoid Robot 892
Prototype, HRP-5P, Capable of Heavy Labor, 2018, 893
https://www.aist.go.jp/aist_e/list/latest_research/2018/20181116/en20181116.html, 894
Accessed date: 5 January 2019. 895
[20] T. Linner, Automated and Robotic Construction: Integrated Automated Construction 896
Sites. Doctoral dissertation, Chair for Building Realization and Robotics, Technische 897
Universität München, 2013, https://mediatum.ub.tum.de/?id=1131018, Accessed date: 898
1 November 2017. 899
https://doi.org/10.1016/j.autcon.2018.05.002http://doi.org/10.22260/isarc2018/0006https://doi.org/10.1016/j.autcon.2018.05.005https://mediatum.ub.tum.de/?id=1131018
-
Accepted manuscript
39
[21] B. Chu, K. Jung, M. T. Lim, D. Hong, Robot-based construction automation: An 900
application to steel beam assembly (Part I), Automation in Construction 32 (2013) 46-901
61. https://doi.org/10.1016/j.autcon.2012.12.016. 902
[22] Y. Hasegawa, T. Onoda, K. Tamaki, Module Robot Application for Flexible Building 903
Construction, in: Proceedings of 9th International Symposium on Automation and 904
Robotics in Construction, 1992, pp. 113–122. 905
https://doi.org/10.22260/ISARC1992/0017. 906
[23] J. Maeda, Development and Application of the SMART System, In: Automation and 907
Robotics in Construction XI, 1994, pp. 457-464, Elsevier Amsterdam. ISBN: 908
0444820442. 909
[24] T. Bock and T. Linner, Robot-Oriented Design: Design and Management Tools for 910
the Deployment of Automation and Robotics in Construction, Cambridge University 911
Press, New York, 2015. ISBN: 9781107076389. 912
[25] T. Linner, W. Pan, R. Hu, C. Zhao, K. Iturralde, M. Taghavi, J. Trummer, M. Schlandt, 913
T. Bock, A Technology Management System for the Development of Single-Task 914
Construction Robots, Construction Innovation 10(1) (2020) 96-111, 915
https://doi.org/10.1108/CI-06-2019-0053. 916
[26] National Aeronautics and Space Administration (NASA), NASA Systems 917
Engineering Handbook. Washington, DC, 2007. ISBN: 9780160797477. 918
[27] A. Warszawski and R. Navon, Implementation of robotics in building: Current status 919
and future prospects, Journal of Construction Engineering and Management 124(1) 920
(1998) 31-41, http://dx.doi.org/10.1061/(ASCE)0733-9364(1998)124:1(31). 921
[28] G. Quezada, A. Bratanova, N. Boughen, S. Hajkowicz, Farsight for construction: 922
Exploratory scenarios for Queensland’s construction industry to 2036, The 923
https://doi.org/10.1016/j.autcon.2012.12.016
-
Accepted manuscript
40
Commonwealth Scientific and Industrial Research Organisation, Australia, 2016. 924
https://doi.org/10.4225/08/58557e0b380ab. 925
[29] Hong Kong Trade Development Council (HKTDC), Real Estate and Construction 926
Services Industry in Hong Kong, HKTDC, Hong Kong, 2018. http://hong-kong-927
economy-research.hktdc.com/business-news/article/Hong-Kong-Industry-928
Profiles/Real-Estate-and-Construction-Services-Industry-in-Hong-929
Kong/hkip/en/1/1X000000/1X003UNV.htm, Accessed date: 1 October 2018. 930
[30] W. Pan and M. Pan, A dialectical system framework of zero carbon emission building 931
policy for high-rise high-density cities: Perspectives from Hong Kong, Journal of 932
Cleaner Production 205 (2018) 1-13. https://doi.org/10.1016/j.jclepro.2018.09.025. 933
[31] Construction Industry Council (CIC), Report of CIC Manpower Forecasting Model 934
2018 (Skilled Construction Workers), CIC, Hong Kong, 2019, 935
http://www.cic.hk/files/page/56/Manpower%20Forecast%20-936
%20Workers%20(2019-2023)_e_20190503.pdf, Accessed date: 1 July 2019. 937
[32] Rider Levett Bucknall (RLB), Hong Kong Report-Quarterly Construction Cost 938
Update Q1 2017, RLB, 2017. https://www.rlb.com/wp-939
content/uploads/2017/04/RLB-Construction-Cost-Update-HK-Q1-2017.pdf, 940
Accessed date: 30 May 2018. 941
[33] Hong Kong Green Building Council (HKGBC), Market drivers for transformation 942
of green buildings in Hong Kong (Executive Summary), HKGBC, Hong Kong, 2014. 943
https://www2.hkgbc.org.hk/ebook/HKGBC_Roadmap/files/assets/basic-html/page-944
1.html, Accessed date: 2 May 2017. 945
https://doi.org/10.1016/j.jclepro.2018.09.025
-
Accepted manuscript
41
[34] W. Pan and C. K. Hon, Modular integrated construction for high-rise buildings, In: 946
Proceedings of The Institute of Civil Engineers: Municipal Engineer. 947
https://doi.org/10.1680/jmuen.18.00028. 948
[35] T. Bock, W. Pan, R. Hu, T. Linner, Investigating the Potential of Implementing 949
Robotics and Automation in the Context of Large-scale Housing Development for 950
Hong Kong, Consultancy Report, Construction Industry Council, Hong Kong. 951
http://www.cic.hk/files/page/56/CIC_robotics%20study_final%20report.pdf, 952
Accessed date: 1 July 2019. 953
[36] M. Pan, and W. Pan, Determinants of adoption of robotics in precast concrete 954
production for buildings, Journal of Management in Engineering 35(5) (2019) 955
05019007. https://doi.org/10.1061/(ASCE)ME.1943-5479.0000706. 956
[37] Y. Yang, M. Pan, W. Pan, Co-evolution through interaction’of innovative building 957
technologies: The case of modular integrated construction and robotics, Automation 958
in Construction 107 (2019) 102932. https://doi.org/10.1016/j.autcon.2019.102932. 959
[38] Gammon, The rise of robotics Gammon technologies and the changing face of 960
construction, The Record, 1 (2017) 12-17. 961
https://www.gammonconstruction.com/uploaded_files/ 962
publication/1/94/Issue%201%202017%20FINAL.pdf, Accessed date: 10 May 2018. 963
[39] Construction Industry Council (CIC), Construction Innovation and Technology 964
Application Centre, CIC, Hong Kong, 2017, http://www.cic.hk/eng/main/ 965
innotechcentre, Accessed date: 1 March 2018. 966
[40] Planning Department, Hong Kong 2030+: Towards a Planning Vision and Strategy 967
Transcending 2030. Planning Department, Hong Kong, 2016. 968
https://doi.org/10.1680/jmuen.18.00028https://doi.org/10.1016/j.autcon.2019.102932
-
Accepted manuscript
42
https://www.hk2030plus.hk/document/2030+Booklet_Eng.pdf, Accessed date: 1 969
March 2018. 970
[41] P. Bishop, A. Hines, T. Collins, The current state of scenario development: an 971
overview of techniques, Foresight 9(1) (2007) 5-25. 972
https://doi.org/10.1108/14636680710727516. 973
[42] B. Stelzer, F. Meyer-Brötz, E. Schiebel, L. Brecht, Combining the scenario technique 974
with bibliometrics for technology foresight: The case of personalized medicine, 975
Technological Forecasting and Social Change 98 (2015) 137-156. 976
https://doi.org/10.1016/j.techfore.2015.06.008. 977
[43] M. Amer, T. U. Daim, A. Jetter, A review of scenario planning, Futures 46 (2013), 978
23-40. https://doi.org/10.1016/j.futures.2012.10.003. 979
[44] D. Mietzner and G. Reger, Advantages and disadvantages of scenario approaches for 980
strategic foresight, International Journal of Technology Intelligence and Planning 1(2) 981
(2005), 220-239. https://doi.org/10.1504/IJTIP.2005.006516. 982
[45] T. U. Daim, G. Rueda, H. Martin, P. Gerdsri, Forecasting emerging technologies: 983
Use of bibliometrics and patent analysis, Technological Forecasting and Social 984
Change 73(8) (2006) 981-1012. https://doi.org/10.1016/j.techfore.2006.04.004. 985
[46] Y. Geum, S. Lee, Y. Park, Combining technology roadmap and system dynamics 986
simulation to support scenario-planning: A case of car-sharing service, Computers and 987
Industrial Engineering, 71 (2014) 37-49. https://doi.org/10.1016/j.cie.2014.02.007. 988
[47] B. Mazzorana and S. Fuchs, Fuzzy formative scenario analysis for woody material 989
transport related risks in mountain torrents. Environmental Modelling & Software, 990
25(10) (2010) 1208-1224. https://doi.org/10.1016/j.envsoft.2010.03.030. 991
https://doi.org/10.1016/j.techfore.2006.04.004
-
Accepted manuscript
43
[48] M. Lettner, J. P. Schöggl, T. Stern, Factors influencing the market diffusion of bio-992
based plastics: Results of four comparative scenario analyses, Journal of Cleaner 993
Production 157 (2017) 289-298. https://doi.org/10.1016/j.jclepro.2017.04.077. 994
[49] J. I. Shieh, H. H. Wu, K. K. Huang, A DEMATEL method in identifying key success 995
factors of hospital service quality, Knowledge-Based Systems 2